What Latvia has done nationally under Regulation (EU) 2024/1689 as amended by Regulation (EU) 2026/1744: the accompanying law, which bodies it has designated as market surveillance authority, single point of contact and notifying authority, its Article 99 penalty regime, whether an Article 57 regulatory sandbox exists (the deadline is 2 August 2027), and any national guidance that reaches courts, law firms or legal-AI deployers. Each row is dated and sourced to the gazette, parliament or authority; titles in the original language are quoted, English titles marked where translated by us.
Informative report on implementing the AI Act requirementstranslated
Informatīvais ziņojums "Par Mākslīgā intelekta akta prasību ieviešanu"
Latvia's TAP portal publishes a Cabinet protocol-decision draft and informative report for organising national AI Act implementation. The draft records proposed authority roles, staffing needs from 2026 and responsibility allocation for EU-level communication and cooperation.
Authority: Ekonomikas ministrija; Latvijas nacionālais akreditācijas birojs — notifying authority
The Latvian Cabinet draft assigns AI Act notifying-authority functions to the Ministry of Economics. It also records that the Latvian National Accreditation Bureau will carry out national accreditation-body functions for AI Act conformity-assessment work.
“"EM veic MI aktā noteiktās paziņojošās iestādes funkcijas"”
Authority: Patērētāju tiesību aizsardzības centrs and listed sector authorities — market surveillance authority, single point of contact
The Latvian draft lists PTAC and sector bodies as market surveillance authorities and names PTAC as the single contact point for communication with the public, Member State partners and EU institutions under the AI Act.
“"Noteikt PTAC par MI akta vienoto kontaktpunktu"”
Authority: Datu valsts inspekcija — market surveillance authority
The Latvian draft assigns the Data State Inspectorate market-surveillance tasks for specified Annex III high-risk AI systems, including law-enforcement, border-management, justice and democracy uses, and supervision of prohibited AI practices under Article 5.
“"DVI nodrošināt MI tirgus uzraudzību"”
Regulatory sandbox (Art. 57)
Regulatory sandbox (Art. 57) · 2026-01-13 · announced
VARAM develops rules for safe testing of AI solutions in a special regulatory environmenttranslated
Authority: Mākslīgā intelekta centrs — sandbox operator
VARAM reports Cabinet approval of rules for the Artificial Intelligence Centre to organise a special regulatory environment for AI testing. The page describes supervised projects, consultations, data safeguards and a planned first call.
Cabinet-approved rules for a special regulatory environment; the first project call was planned for February 2026.
“"Speciālā regulatīvā vide izveidota kā kontrolēts un drošs ietvars"”
Authorities in the EU dataset
Patērētāju tiesību aizsardzības centrs— single-point-of-contactsource